Hi all;

I've been looking for periodicities in small quakes at specific locations.

I do a search then sort the quakes into a big array where each cell is a day starting with the first quake.

Then I divide the big array up into a small array whose size is the period I'm examining.

Does this make sense so far?

By doing this repeatedly with larger and larger arrays I can examine periods from 5 to 60 days by plotting the cell totals on a graph.

I'm looking for periods with a single big peak.

If you set a prediction window there, you can catch more quakes than chance would allow.

Does anyone know of a better way to do this?

Roger

PS: There's no assurance the pattern will continue into the future but it's a start.